Reference-led generation
Separate identity, place and movement
Let a portrait own identity, a location image own the space and a video reference own only camera movement.
Kling Video O1
Unify text, boundary frames, references and semantic edits in one video brief.
Multimodal console
One model from generation to revision
Kling Video O1 treats text, images, video and named subjects as instructions inside one creation system. In Flux Art, choose the input route first, then state what every source owns and what may change.
Unified multimodal inputUp to 7 image referencesVideo reference and editingStandard and Pro
Semantic video editing
Change one visual layer at a time
Keep product, framing and camera move while revising only light or a distracting object.
Boundary frames
Lock the opening and ending before motion
Let two boundary images own the result while the prompt supplies path, camera, physics and continuity.
Input routing
Decide which input owns the information
The input route defines existing visual facts. The prompt should supply motion and change instead of competing with the source.
Text to video
Begin with a complete director brief and no visual source.
First frame
The frame owns subject and opening composition; the prompt owns motion.
Start and end
Two images define the boundaries; the prompt describes the transition.
Multimodal refs
Assign up to seven images and one video to people, objects, places or motion.
Video editing
Edit a 3–10 second source clip and choose whether to keep its sound.
Capability matrix
Where Kling Video O1 fits
Use O1 when the task moves between creation and revision, or when identity, objects, setting and motion must come from separate sources rather than one prompt.
Reference-led generation
Assign people, products, locations and visual treatment to separate images, with an optional video reference for movement or temporal context.
Start and end boundaries
Use two images to define the opening and closing states, then describe only the action, camera move and continuity required between them.
Semantic video changes
Describe a narrow change such as removing a passerby, changing time of day or replacing clothing without drawing masks or animation keyframes.
Subject continuity
Name the identity, garment, prop and environment details that must survive camera movement and multi-subject interaction.
Edit directions
Kling Video O1 director briefs
A useful O1 prompt reads like a compact edit decision list: source roles, desired change, temporal action and protected details.
Documentary arrival
Reference 1 owns the filmmaker's face and saffron raincoat. Reference 2 owns the rain-soaked tram stop. Use the uploaded motion clip only for the slow shoulder-level tracking move. She raises the camera as the tram arrives; preserve her identity, clothing, camera body and rainy night lighting. No readable signs.
Product light revision
Keep the uploaded clip's projector geometry, plinth, composition and camera speed. Change only the lighting: move one restrained cobalt beam from left to right and deepen the concrete texture. Remove dust in the foreground. Do not add text, hands or extra products; keep the original sound.
Coastal boundary transition
The first frame owns the distant lighthouse and empty wet road. The last frame owns the silver coupe after it rounds the bend. Between them, the car enters from the lower left, follows the curve with believable tire grip and passes through thin sea mist. Slow lateral camera pan; preserve blue-hour color and road geometry.
Delivery check
How to use Kling Video O1 in Flux Art
Kuaishou introduced Kling O1 on December 1, 2025 as a unified multimodal creation model built on an MVL framework. Flux Art currently exposes text, first-frame, start-and-end-frame, multimodal reference and existing-video editing routes with Standard and Pro modes and prompt-adherence control.
Create with Kling Video O1- Choose the input contractStart from text, one frame, two boundary frames, multimodal references or an existing 3–10 second clip. Do not upload a source without assigning it a job.
- Write change and keep rulesDescribe the event in time order, then list identity, geometry, framing, camera motion or sound that must remain unchanged.
- Generate and review by layerChoose Standard or Pro, set adherence, then review subject identity, object structure, motion, edit boundary and continuity separately.
Source notes
Official sources and page information
Model capabilities are summarized from public Kling AI and Kuaishou materials. Available modes, parameters and pricing follow the current Flux Art workspace.
Published
Compare related Kling video models
- Kling V3 OmniDirect text, frames, images and video references as one 3–15 second audiovisual brief.
- Kling V3Direct image, movement, dialogue and ambience in one flexible 3–15 second shot plan.
- Kling V2.6Generate image, Chinese or English speech, effects and ambience together in clips up to 10 seconds.
- Kling Image O1Direct one image with text and up to ten visual references.
Kling Video O1 FAQ
What is Kling Video O1?
Kling Video O1 is a unified multimodal video model announced by Kuaishou in December 2025. It brings text, images, video clips and subject references into one system for generation and semantic editing.
Which Kling Video O1 workflows are available in Flux Art?
Flux Art currently provides text-to-video, first-frame, start-and-end-frame, multimodal reference and existing-video editing routes for Kling Video O1.
How many reference files can I use with Kling Video O1?
The current Flux Art multimodal route accepts up to seven images and one video reference. Give every file a specific role such as person, product, place, style or movement.
Can Kling Video O1 use start and end frames?
Yes. Upload two boundary images and describe only the action, camera movement and continuity needed between the opening and closing states.
Can Kling Video O1 edit an existing video?
Yes. The Flux Art editing route accepts a 3–10 second source clip and lets you choose whether to keep its original sound. Narrow, explicit changes are easier to review than unrestricted restyling.
Which output modes does Flux Art expose for Kling Video O1?
The current integration exposes Standard and Pro modes plus prompt-adherence control. Available controls in the generation workspace are the source of truth if the provider changes its options.
How should I write a multimodal O1 prompt?
Name every input, state what it owns, write the event in time order, then list the identity, geometry, framing, camera motion or sound that must remain unchanged.
How do I improve character or product consistency?
Use clear reference views and describe the exact facial, clothing, shape, label, material and color details that must be preserved. Avoid giving multiple sources responsibility for the same feature.
Is Kling Video O1 the same as Kling V3 Omni?
No. Kling's official Video 3.0 guide describes Video 3.0 Omni as the upgrade path from Video O1. O1 remains a separate model and should not be credited with later 3.0 Omni features unless the current workspace exposes them.
Can Kling Video O1 results be used commercially?
Commercial use depends on the current Flux Art and model-provider terms and on your rights to prompts and reference media. Check the latest terms and obtain necessary permissions before client or advertising use.